Review Taylor Series and Error Analysis Roots of Equations Linear Algebraic Equations Optimization...

Post on 12-Jan-2016

221 views 4 download

Transcript of Review Taylor Series and Error Analysis Roots of Equations Linear Algebraic Equations Optimization...

Review

• Taylor Series and Error Analysis• Roots of Equations• Linear Algebraic Equations• Optimization• Numerical Differentiation and Integration• Ordinary Differential Equations• Partial Differential Equations• Curve Fitting

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

1

Taylor Series

• Lagrange remainder

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

2

Roots of Equations

• Bracketing Methods• Bisection Method

• False Position Method

• Open Methods• Fixed Point Iteration

• Newton-Raphson Method

• Secant Method

• Roots of Polynomials• Müller’s Method

• Bairstow’s Method

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

3

Bisection Method

• Example:

• Use range of [202:204]

• Root is in upper subinterval

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

4

Bisection Method

• Use range of [203:204]

• Root is in lower subintervalNumerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

5

Fixed Point Iteration Example

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

6

Special attention

Read Chap 6.1, 6.6

Newton-Raphson Method

• Use tangent to guide you to the root

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

7

Linear Algebraic Systems

• Gaussian Elimination• Forward Elimination• Back Substitution

• LU Decomposition

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

8

Gaussian Elimination

• Forward elimination

• Eliminate x1 from row 2

• Multiply row 1 by a21/a11

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

9

Gaussian Elimination

• Eliminate x1 from row 2• Subtract row 1 from row 2

• Eliminate x1 from all other rows in the same way

• Then eliminate x2 from rows 3-n and so onNumerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

10

Gaussian Elimination

• Forward elimination

• Back substitute to solve for x

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

11

LU Decomposition

• Substitute the factorization into the linear system

• We have transformed the problem into two steps• Factorize A into L and U• Solve the two sub-problems

• LD = B• UX = D

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

12

LU Decomposition

• Example

• Factorize A using forward elimination

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

13

LU Decomposition

• Example

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

14

LU Decomposition

• Example

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

15

LU Decomposition

• Example

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

16

Optimization Methods• One-dimensional unconstrained optimization

• Golden-Section

• Quadratic Interpolation

• Newton’s Method

• Multidimensional unconstrained optimization• Direct Methods

• Gradient Methods

• Constrained Optimization• Linear Programming

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

17

Golden-section search

• Algorithm• Pick two interior points in the interval using the

golden ratio

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

18

Golden-section search

• Two possibilities

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

19

Golden-section search

• Example

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

20

Golden-section search

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

21

Golden-section search

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

22

Newton’s Method

• Newton-Raphson could be used to find the root of an function

• When finding a function optimum, use the fact that we want to find the root of the derivative and apply Newton-Raphson

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

23

Newton’s Method

• Example

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

24

Newton’s Method

• Example

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

25

Quadratic interpolation

• Use a second order polynomial as an approximation of the function near the optimum

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

26

Special attention

Gradient Methods

• Given a starting point, use the gradient to tell you which direction to proceed

• The gradient gives you the largest slope out from the current position

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

27

Special attention

Numerical Integration• Newton-Cotes

• Trapezoidal Rule• Simpson’s Rules (Special attention for

unevenly distributed points)

• Romberg Integration• Gauss Quadrature

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

28

Newton-Cotes Formulas• Trapezoidal Rule

• Simpson’s 1/3 Rule

• Simpson’s 3/8 Rule

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

29

Special attention

Read Chap 21.2-3

Integration of Equations

• Romberg Integration• Use two estimates of integration and then

extrapolate to get a better estimate

• Special case where you always halve the interval - i.e. h2=h1/2

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

30

Romberg Integration

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

31

Ordinary Differential Equations• Runge-Kutta Methods

• Euler’s Method• Heun’s Method• RK4

• Multistep Methods• Boundary Value Problems• Eigenvalue Problems

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

32

Euler’s Method

• Example:

• True:• h=0.5

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

33

Heun’s Method

• Local truncation error is O(h3) and global truncation error is O(h2)

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

34

Heun’s Method

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

35

Classic 4th-order R-K method

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

36

Special attention to

ODE equation system

Not only one equation

Curve Fitting

• Least Squares Regression• Interpolation• Fourier Approximation

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

37

Polynomial Regression• An mth order polynomial will require that you

solve a system of m+1 linear equations

• Standard error

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

38

Special attention

Lecture note 19

Chap 17.1

Newton (divided difference) Interpolation polynomials

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

39

Newton (divided difference) Interpolation polynomials

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

40

Interpolation• General Scheme for Divided Difference

Coefficients

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

41

Interpolation

• General Scheme for Divided Difference Coefficients

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

42

Interpolation

• Example:• Estimate ln 2 with data points at (1,0),

(4,1.386294)• Linear interpolation

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

43

Interpolation

• Example:• Estimate ln 2 with data points at (1,0),

(4,1.386294), (5,1.609438)• Quadratic interpolation

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

44

Interpolation

• Example:• Estimate ln 2 with data points at (1,0),

(4,1.386294), (5,1.609438), (6,1.791759)• Cubic interpolation

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

45

Spline Interpolation

• Spline interpolation applies low-order polynomial to connect two neighboring points and uses it to interpolate between them.

• Typical Spline functions

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

46

Cubic Spline Functions

• This gives us n-1 equation with n-1 unknowns – the second derivatives

• Once we solve for the second derivatives, we can plug it into the Lagrange interpolating polynomial for second derivative to solve for the splines

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

47

Cubic Spline Functions

• Example: (3,2.5), (4.5,1), (7,2.5), (9,0.5)• At x=x1=4

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

48

Cubic Spline Functions

• At x=x2=7

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

49

Cubic Spline Functions

• Solve the system of equations to find the second derivatives

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

50

Cubic Spline Equations

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

51

Cubic Spline Equations

• Substituting for other intervals

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

52

Final Exam• December 13 Friday, 10:30 AM~12:30

PM, ITE 119• Closed book, three cheat sheets

(8.5x11in) allowed • Office hours:

• December 12, 1-3pm, or by appointment• TA December 10, 11am-12noon or by

appointment

Numerical MethodsLecture 22

Prof. Jinbo BiCSE, UConn

53